Prime Video international head Kelly Day has outlined how the streamer is applying artificial intelligence across production and audience services. Speaking at the RTS Convention, she identified previsualisation, sound design and visual-effects workflows as areas that can run in parallel and reduce repetitive work.
On the viewer side, Prime Video already uses AI for recommendations, simplified synopses and spoiler-controlled summaries. Dialogue Boost raises speech relative to background sound, while the service is expanding audio descriptions and subtitles. A recent Maxton Hall workflow synchronized actors’ lip movements to human-performed English dubbing.
The practical distinction is between automation around a creative decision and automated authorship. Prime Video says it is not seeking a model to write a programme, but tools that reduce administration, shorten iterative tasks and increase accessibility. Day also described guardrails as a foundation rather than a later compliance layer.
The speech confirms corporate direction and named examples, but it does not publish models, training data, consent processes, error rates, accessibility quality measurements or the review workflow for altered faces and voices. Cost and labour effects also remain unquantified.
